[Publication Time] 2021.09.01

[Lead Author] Keliang Chang

[Corresponding Author] Chen Guijing

[Journal] ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH

[Abstract]

The Logarithmic Mean Divisia Index (LMDI) model is applied to study Chinese national and regional power sector carbon emission changes through consumption side from 2003 to 2017, and regional power sector carbon emissions are estimated through the production and consumption accounting principle. The two-factor ANOVA and one-factor ANOVA are used to compare the differences of regional power sector carbon emissions through the two principles. In addition, the Tapio decoupling analysis model is used to investigate the decoupling state between carbon emissions of power sector and the corresponding driving forces through the consumption side. There are several results: (1) Through the two different principles, regional power sector carbon emissions are statistically significant, yet national power sector carbon emissions are not statistically significant; (2) the main factors contributing to the power sector carbon emission growth are economic scale effect and income level effect, and the main restraining factors are electricity consumption carbon intensity effect and production sector electricity intensity effect; (3) the highest contribution effect to the decoupling indexes between various influencing factors and power sector carbon emissions was scale effect, and technical effect had the second largest contribution value; (4) in 2003-2017, economic scale effect was the first significant factor causing the difference of regional power sector carbon emissions, followed by production sector electricity intensity effect and electricity consumption carbon intensity through the regional decomposition analysis. Finally, this paper gives some targeted suggestions for the low-carbon development of the power sector through national and regional perspectives.

[Keywords]

Carbon emissions; China; Consumption accounting principle; LMDI; Power sector; Tapio decoupling analysis model